Statistical predictor identification

By Hirotugu Akaike. 1970

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Deep learning in neural networks: An overview (2014, background): “…sets (Craven & Wahba, 1979; Eubank, 1988; Golub, Heath, & Wahba, 1979; Hastie & Tibshirani, 1990; Mosteller & Tukey, 1968; Stone, 1974), Akaike’s information criterion and final prediction error (Akaike, 1970, 1973, 1974), or generalized prediction error (Moody, 1992; Moody & Utans, 1994).”
On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models (2015, background): “2 M ’s Training To decrease bitsM + bitsH , we add a regularizing term to E, to punish excessive complexity [4, 5, 91, 294, 155, 135, 97, 170, 169, 104, 290, 7, 290, 87, 286, 319, 100, 102].”
Regression model selection—a residual likelihood approach (2002, methodology): “Selection criteria are usually classified into one of two categories on the basis of their second element: efficient (e.g. AIC, AICC, FPE and Cp) or consistent (e.g. BIC).”
Jackknife model averaging for quantile regressions (2015, unknown): “For example, Hurwich and Tsai (1990) develop a small sample criterion for the selection of LAD regression models; Machado (1993) and Burman and Nolan (1995) consider variants of the Schwarz information criterion (SIC) and Akaike information criterion (AIC), respectively, for -estimation which inclu”
Variable Selection in Predictive Regressions (2013, methodology): “A p-th order autoregression uses p lags and the model selection problem reduces to the determination of the lag length, p. Akaike (1969, 1970) propose to measure adequacy by the final prediction error E (yT+h − ŷT+h)2 which can be viewed as a weighted version of Mallows’ criterion with all weight…”
Economic growth and tax components: an analysis of tax changes in OECD (2005, methodology): “Armah following Hsiao ( 1979 , 1981 , 1982 ) overcomes this problem by combining Akaike's ( 1969 and 1970 ) Final Prediction Error (FPE) criterion with Granger's causality test.”
Model Selection for Small Sample Regression (2002, background): “Note that in the asymptotic case, since the covariance matrix is almost the identity matrix (from the orthonormality assumption), E(1/λ i ) ≈ 1 and we obtain Akaike's term (4).”
The Focused Information Criterion (2003, background): “Nishii (1984) showed that AIC, nite prediction error (Akaike 1970), Mallows’s (1973) Cp, and the prediction sum of squares (Allen 1971) are equivalent in having the same risk function in the limit.”
A survey of cross-validation procedures for model selection (2009, unknown): “Examples of model selection procedures following the unbiased risk estimation principle are FPE (Final Prediction Error, Akaike, 1970), several cross-validation procedures including the Leave-one-out (see Section 4), and GCV (Generalized Cross-Validation, Craven and Wahba, 1979, see Section 4.3.3).”
Blind detection of independent dynamic components (2001, methodology): “In figure 2 we show the value of a “cost function” defined as the average negative log likelihood log P (Xjm), estimated on a test set, by Akaike’s information criterion (AIC) [13], and by”